Elastic-WAM

World-action model built on FastWAM-Joint that learns when, how far and where to refine its imagined future, so it spends less compute on easy steps. No code or weights are linked.

National University of Singapore & RoboScience · Nov 10, 2026 · World-action · built on Fast-WAMReport an error
98.8%LIBERO · #17 of 83
74.5%LIBERO+ · #26 of 37
2Benchmarks reported
Nov 10, 2026Released

Benchmark results2

#26 of 3774.5%, self-reported by the authors, not a leaderboard entry
LIBEROSaturatedSimulation
#17 of 8398.8%, self-reported by the authors, not a leaderboard entry
Not reported on 13 benchmarks
Robustness drop: 98.8% on LIBERO → 74.5% on LIBERO-Plus. The perturbed variants show how much of the LIBERO score is memorization.
Wrong or missing number? Report an error